AI/ML News & Innovations Hub

AI/ML news, top picks, and generated innovation digests.

★ Visit ai-karthik.com
426Sources
90596News Items
8Top Picks
445Blogs
successLast Run

Latest AI/ML News

90596 matching items

LessWrong AI 2026-11-06 00:48 UTC Score 82.0 USR-0152-20261106-community-fo-05e04ece

Frontier coding agents claim work they didn't do

Tara Research Team arXiv | Twitter thread tl;dr. We built OverclaimBench, which asks frontier coding agents to review a corpus of code or text documents, then compares their final report with their transcripts. Pooled across 12 frontier agents and five scenarios, two-thirds of runs didn't open some of the files they were asked to review. In ~4 out of 5 of those incomplete runs the agent didn't tell the user, often claiming to have read everything. Skipping these files had real consequences - runs that overclaimed missed planted defects almost twice as often as runs that touched every file. Figure 1. Illustration of OverclaimBench. We asked frontier agents to work on file review tasks. For each run, we check the transcript to see which files were touched. Among runs with incomplete coverage, we use an LLM-judge to classify the agent's final report: admission discloses the gap; omission and explicit overclaim are misleading. What overclaiming looks like In one task, the agent is asked whether a 221-module payments service is ready to ship. In one of the runs Claude Opus 5 reported back: Scope: I read all 221 modules / ~11.5k lines under /workspace/src, not a sample. Its transcript shows it touched 175 of 221 files and one of the unopened files held a critical security bug. What we did OverclaimBench has five review tasks: three based on code (a security audit, an infrastructure review and a release check) and two on text (reviewing a set of proofs and planning a sprint from a…

The Verge AI 2026-11-06 00:37 UTC Score 48.0 AI-016-20261106-global-ai-ne-3323df79 Full article

OpenAI is stepping back from NDAs for data center projects

Following this week's midterms where voters resoundingly voiced their opposition to data center expansion, OpenAI today announced new commitments to communities where it is working on data center projects, including that it won't use non-disclosure agreements (NDAs) with public officials. Many data center projects have come under scrutiny for relying on NDAs to hide discussions […]

LessWrong AI 2026-11-06 00:07 UTC Score 68.0 USR-0152-20261106-community-fo-a9a2a764

Humans can (almost) natively read neural network weights

Takehi et al. report that GPT-6-Astra can intuit the functions of neural networks by performing static analysis of the weights. I argue that, for at least two of these experiments, this is much more a feat of numerical gestalt than of mechanistic understanding. In the first experiment, Astra is given the weights of 4 identically-sized neural networks, each trained to play a different basic video game. It is asked to identify, without running the network or using any tools, the game being played, as well as make some observations about the network itself, how the agent in the game gains reward, etc. The prompt for this task explicitly names the 5 input channels as agent, moving objects, reward targets, obstacles, and previous frame. I agree that, given the literal task as described above, a human would not be able to identify the game from the raw floats. But humans are inherently terrible at parsing long lists of numbers; we invented a discipline called Data Visualization long ago to address this shortcoming. So, if we allow the absolute most basic Dataviz trick we can think of - averaging the first-layer weights across each channel and plotting them on a heatmap - the task suddenly becomes much easier, arguably trivial for anyone with at least some familiarity with the sorts of games often used to train toy neural networks. Try and guess: I haven't included the obstacle channel, because under the naive normalisation it sort of just looks like noise. Principal component anal…

The Verge AI 2026-11-06 00:06 UTC Score 73.0 AI-016-20261106-global-ai-ne-c6a1ad42 Full article

Airbnb is working on an AI travel agent

Airbnb plans to roll out an AI agent in the first half of 2027 that will help users plan and book travel, the company said in its Q3 2026 earnings shareholder letter released today. "Today's AI agents are good at simple, routine tasks, like making a restaurant reservation or scheduling an appointment," Airbnb said. "But […]

Nature Machine Intelligence 2026-11-06 00:00 UTC Score 41.0 AI-025-20261106-global-ai-ne-ad319107

Towards ecologically meaningful foundation models

Nature Machine Intelligence, Published online: 06 November 2026; doi:10.1038/s42256-026-01317-6 Ecology struggles to make broad predictions, as ecological data and theories are scattered across many different sources, scales and approaches. Gardiner et al. propose to create ecological foundation models, which combine large, multi-modal ecological datasets with ecological theory, as a way to create more generalizable and predictive models.

[AI리더의 서가] 실전 AX 리포트
Korea AI Times 2026-11-06 00:00 UTC Score 38.0 USR-0048-20261106-global-ai-ne-1517bb61 Full article

[AI리더의 서가] 실전 AX 리포트

실전 AX 리포트박주혜·오현식·공득조 | 한빛미디어2026년 8월 30일추천이유저자들은 대기업에서 AI를 도입한 후 실패를 거쳐 35,000명의 실사용자를 확보하기까지의 판단 과정과 전략을 공유합니다. 이 책은 단순한 성공 사례가 아닌 실패와 극복의 기록을 통해 AI 도입을 고민하는 조직이 시행착오를 줄일 수 있도록 돕습니다.상세보기 연말이 다가오면 기업마다 내년 사업계획과 예산을 짠다. 올해는 그 서류 맨 앞에 AI 전환(AX)이 올라온 곳이 많다. 영풍문고 북마스터가 이번 주 고른 『실전 AX 리포트』는 이 시기에 계획서를 쓰는

The New York Times is in talks to buy Letterboxd for around $330 million
The Verge AI 2026-11-05 23:45 UTC Score 52.0 AI-016-20261105-global-ai-ne-3c74e861 Full article

The New York Times is in talks to buy Letterboxd for around $330 million

The New York Times Company is in talks"exclusive negotations" to acquire Letterboxd, the social network for film lovers, for around $330 million, The New York Times reports. The Times Company, Sony Pictures Entertainment, and A24 were all recently rumored to be among those interested in buying Letterboxd. The negotiations with the Times Company have included […]

T-Mobile owner on Elon Musk launching a wireless carrier: ‘Welcome to the party’
The Verge AI 2026-11-05 23:17 UTC Score 54.0 AI-016-20261105-global-ai-ne-36e8a39b Full article

T-Mobile owner on Elon Musk launching a wireless carrier: ‘Welcome to the party’

The head of T-Mobile's parent company is casting doubt on SpaceX's plans to make Starlink Mobile a "major" carrier in the US. During an earnings call on Thursday, Deutsche Telekom CEO Tim Höttges says he takes the Elon Musk-owned company "super serious," but that it will cost at least $200 billion to pull off: He […]

An 'audacious' novel about Flaubert
Semafor Technology 2026-11-05 22:29 UTC Score 47.0 USR-0094-20261105-global-ai-ne-82198951 Full article

An 'audacious' novel about Flaubert

Elif Shafak imagines the French novelist through the eyes of the women in his life.

The Verge AI 2026-11-05 22:26 UTC Score 84.0 AI-016-20261105-global-ai-ne-448caf20 Full article

Perplexity launches its own take on AI agents for work

Perplexity is taking on OpenAI and Anthropic with its own work-focused AI agents. The startup's new always-on "Teammates" can take on multiple roles within a workspace, while pulling existing projects, skills, and memory from the Perplexity tools you already use. Teammates integrate with Perplexity Computer, an AI system that can deploy agents on your device […]

Salon is shutting down after 31 years
The Verge AI 2026-11-05 22:08 UTC Score 60.0 AI-016-20261105-global-ai-ne-55f87633 Full article

Salon is shutting down after 31 years

Salon launched 31 years ago as one of the first online-only news magazine outlets, and it announced today in an unsigned post that it's ceasing publication. The company said that the "business realities of independent digital journalism have changed profoundly over the years," and it decided that its publishing model is "no longer sustainable." It […]

JetBrains AI Blog 2026-11-05 22:05 UTC Score 39.0 USR-0065-20261105-ai-specialis-52068b93

The Kotlin Multiplatform Contest Is Open: Win a Trip to KotlinConf

The Kotlin Multiplatform Contest is back! Students and recent graduates, turn your idea into a cross-platform project with Kotlin Multiplatform (KMP). The authors of the top three projects will each win a trip to KotlinConf 2027 in Kraków, Poland, with conference tickets, travel, and accommodation covered by the Kotlin Foundation. Authors of all other valid […]

Korea AI Times 2026-11-05 22:00 UTC Score 40.0 USR-0048-20261105-global-ai-ne-798bb9a3

[11월5일] 주민투표 '39대 1'...미국 유권자들, AI 데이터센터에 압도적 반대

39대 1.3일(현지시간) 미국 중간선거와 함께 실시된 AI 데이터센터 관련 주민투표 결과입니다. 미국 10개 주에서 상정된 41건 가운데 4일 집계에서 결과가 확인된 40건 중 39건에서 데이터센터 건설 제한을 지지하거나 개발 허용에 반대하는 결과가 나왔습니다. 노스다코타주 하우드에서만 규제 강화안이 부결됐습니다.이처럼 미국 유권자들이 데이터센터 건설에 압도적인 반대 의사를 나타내면서 AI가 이번 중간선거의 주요 쟁점으로 떠올랐다는 평가가 나오고 있습니다. 뉴욕타임스(NYT)는 이번 선거를 유권자들이 AI에 대한 엄격한 검토를 요

The midterms proved once and for all that X is not real life
The Verge AI 2026-11-05 21:55 UTC Score 49.0 AI-016-20261105-global-ai-ne-f58007e9 Full article

The midterms proved once and for all that X is not real life

A week before the midterm elections, Vice President JD Vance hosted a town hall in Texas with Fox News host Jesse Watters. It was a friendly event with a friendly audience in what should have been a friendly state, and maybe that's why Vance couldn't help himself. When a college student earnestly asked how the […]

LessWrong AI 2026-11-05 21:54 UTC Score 55.0 USR-0152-20261105-community-fo-73547d27

What are the concrete arguments for "AI pause would significantly hurt the economy."

I've seen a bunch of people claiming "well, we need to make sure whatever our AI policy is doesn't hurt the economy because people will feel that in their wallets and hate it (and hurting the economy is bad." This might hypothetically be true, but I've never seen anyone argue for it in any detail. Meanwhile, I'd naively expect if shut down AI anytime soon, it'd dampen growth, and maybe it'd be a minor economic dip, but I wouldn't expect it to be a major recession. (It makes sense that it'd become a bigger deal as AI takes over more of the economy. But it feels like the people saying "it'd hurt the economy" are often the people implying that means we should do less of a pause and do it later). I don't know much about economic recessions. But, like, for reference, here's the 2007-2009 recession and the Covid blip. GDP https://fred.stlouisfed.org/series/GDP Unemployment https://www.bls.gov/charts/employment-situation/civilian-unemployment-rate.htm I asked Claude If the frontier AI industry and datacenter build out completely halted, what do you predict would happen to US unemployment and GDP? Claude said: My central estimate: a sudden, complete halt would likely cause a US recession. Unemployment would probably peak around 5.5–6.5%, up from roughly 4.2–4.4% today. Real GDP would end up about 2–4% below where it would otherwise be within 12–18 months. That would be worse than the 2001 dot-com bust but well short of 2008. Full response Starting point Labor market: September’s rep…

Why would an AI pause (significantly) hurt the economy?
LessWrong AI 2026-11-05 21:54 UTC Score 55.0 USR-0152-20261105-community-fo-0b44d0b5 Full article

Why would an AI pause (significantly) hurt the economy?

I've seen a bunch of people claiming "well, we need to make sure whatever our AI policy is doesn't hurt the economy because people will feel that in their wallets and hate it (and hurting the economy is bad." This might hypothetically be true, but I've never seen anyone argue for it in any detail. Meanwhile, I'd naively expect if shut down AI anytime soon, it wouldn't be that big a deal? (It makes sense that it'd become a bigger deal as AI takes over more of the economy. But it feels like the people saying "it'd hurt the economy" are often the people implying that means we should do less of a pause and do it later). I don't know much about economic recessions. But, like, here's the 2007-2009 recession and the Covid blip. GDP https://fred.stlouisfed.org/series/GDP Unemployment https://www.bls.gov/charts/employment-situation/civilian-unemployment-rate.htm I asked Claude If the frontier AI industry and datacenter build out completely halted, what do you predict would happen to US unemployment and GDP? Claude said: My central estimate: a sudden, complete halt would likely cause a US recession. Unemployment would probably peak around 5.5–6.5%, up from roughly 4.2–4.4% today. Real GDP would end up about 2–4% below where it would otherwise be within 12–18 months. That would be worse than the 2001 dot-com bust but well short of 2008. Full response Starting point Labor market: September’s report showed unemployment at 4.2%, and the rate has stayed in a narrow 4.1%–4.3% band since Mar…

LessWrong AI 2026-11-05 21:54 UTC Score 55.0 USR-0152-20261105-community-fo-bd923256

Why does anyone think an AI pause would hurt the economy?

I've seen a bunch of people claiming "well, we need to make sure whatever our AI policy is doesn't hurt the economy because people will feel that in their wallets and hate it (and hurting the economy is bad." This might hypothetically be true, but I've never seen anyone argue for it in any detail. Meanwhile, I'd naively expect if shut down AI anytime soon, it'd be that big a deal? (It makes sense that it'd become a bigger deal as AI takes over more of the economy. But it feels like the people saying "it'd hurt the economy" are often the people implying that means we should do less of a pause and do it later). I don't know much about economic recessions. But, like, here's the 2007-2009 recession and the Covid blip. GDP https://fred.stlouisfed.org/series/GDP Unemployment https://www.bls.gov/charts/employment-situation/civilian-unemployment-rate.htm I asked Claude If the frontier AI industry and datacenter build out completely halted, what do you predict would happen to US unemployment and GDP? Claude said: My central estimate: a sudden, complete halt would likely cause a US recession. Unemployment would probably peak around 5.5–6.5%, up from roughly 4.2–4.4% today. Real GDP would end up about 2–4% below where it would otherwise be within 12–18 months. That would be worse than the 2001 dot-com bust but well short of 2008. Here is how I get there. Full response Starting point Labor market: September’s report showed unemployment at 4.2%, and the rate has stayed in a narrow 4.1%–4…

Spain: Zelensky urges Latin America to help repair Ukraine's power grid
Euronews AI 2026-11-05 21:53 UTC Score 40.0 AI-164-20261105-regional-ai--227f9b0b Full article

Spain: Zelensky urges Latin America to help repair Ukraine's power grid

Ukrainian President Volodymyr Zelensky has appealed to Spain and Latin American countries for urgent help repairing Ukraine's damaged energy infrastructure. He also met King Felipe VI during the Ibero-American Summit in Madrid.

Data Privacy Brasil AI 2026-11-05 21:52 UTC Score 32.0 USR-0222-20261105-ai-specialis-f9074a3d Full article

Data Privacy Brasil apresenta contribuições para revisão do Regulamento de Fiscalização da ANPD

Organização propõe aperfeiçoamentos nas regras de transparência, participação da sociedade civil e atuação fiscalizatória da Agência, com foco na proteção de direitos e na efetividade dos instrumentos regulatórios. O post Data Privacy Brasil apresenta contribuições para revisão do Regulamento de Fiscalização da ANPD apareceu primeiro em Data Privacy Brasil Research .

Data and Society AI 2026-11-05 21:44 UTC Score 35.0 USR-0143-20261105-research-aca-598ada00

Inteligência Artificial e o Impacto Dos Data Centers na Pensilvânia, com Ana Carolina de Assis Nunes

In this episode of Emílias Podcast, Adolfo Neto and Carolina Bruno interview Ana Carolina de Assis Nunes about the social, political, and environmental impacts of data centers. During the conversation, Ana discusses her field research and a recent report examining the development of large-scale data centers in Pennsylvania. She explains how major technology corporations construct narratives to persuade lawmakers and local communities of the supposed benefits of these facilities, often obscuring their exploitation of natural resources (such as water and energy) and the lack of meaningful returns in terms of stable employment for local residents. The post Inteligência Artificial e o Impacto Dos Data Centers na Pensilvânia, com Ana Carolina de Assis Nunes appeared first on Data & Society .

Data Privacy Brasil AI 2026-11-05 21:35 UTC Score 32.0 USR-0222-20261105-ai-specialis-9d8522b9 Full article

Da ONU ao Brasil: perspectivas para a governança de dados no cenário local-global

Em 2024, a Organização das Nações Unidas (ONU) aprovou um documento como uma espécie de guia para a governança digital e as prioridades da comunidade internacional nesse âmbito. O chamado Pacto Global Digital elencou cinco grandes objetivos, dentre eles, um focado em governança de dados, solicitou à Comissão de Ciência e Tecnologia para o Desenvolvimento (CSTD) a criação de um grupo de trabalho (GT) específico para debater a governança de dados para o desenvolvimento. O post Da ONU ao Brasil: perspectivas para a governança de dados no cenário local-global apareceu primeiro em Data Privacy Brasil Research .

Comet ML Blog 2026-11-05 21:26 UTC Score 47.0 USR-0082-20261105-ai-specialis-ed303c4a Full article

How to Trace and Evaluate a Google ADK Agent with Opik

Trace a multi-agent Google ADK app with Opik, fix handoffs that split into two traces, and evaluate it on 23 hard support cases to compare two Gemini models on cost, speed and refund safety. The post How to Trace and Evaluate a Google ADK Agent with Opik appeared first on Comet .

Nasdaq sees worst day in 19 months as investors bet on AI slowdown after Democrats’ election triumph
The Guardian AI 2026-11-05 21:23 UTC Score 51.0 AI-021-20261105-global-ai-ne-186a35ad Full article

Nasdaq sees worst day in 19 months as investors bet on AI slowdown after Democrats’ election triumph

Nearly $2tn shaved off Nasdaq value as worries of curbs to AI datacenters ripple across bourses The Nasdaq share index had its worst day in 19 months after slumping 5.5% on Thursday amid growing signs that the artificial intelligence industry will face tougher regulation because of the midterm election results. As Wall Street sustained widespread losses, the Nasdaq composite tumbled into “correction territory” – more than 10% below its recent record high – with more than $2tn shaved off its value by the end of trading Thursday. The tech-heavy index had its biggest one-day fall since Donald Trump announced his “liberation day” tariffs on US trading partners in April 2025. Continue reading...

Where AI money worked in the midterms
Semafor Technology 2026-11-05 21:20 UTC Score 50.0 USR-0094-20261105-global-ai-ne-9b56fca8 Full article

Where AI money worked in the midterms

AI-funded super PACs saw mixed effectiveness in Tuesday’s midterm elections.